Nutrition 103
104 (2022) 111769

Contents lists available at ScienceDirect

Nutrition
journal homepage: www.nutritionjrnl.com

Applied nutritional investigation

Associations between soft-drink consumption and lifestyle patterns in

overweight and obese European adults: Feel4 Diabetes Study
Kyriakos Reppas M.Sc. a, Maria Michelle Papamichael Ph.D. a,b, Theodora Boutsikou Ph.D. c,
Greet Cardon Ph.D. d, Violeta Iotova Ph.D. e, Tanya Stefanova Ph.D. e, Roumyana Dimova Ph.D. f,
Rurik Imre Ph.D. g, Anett Sandor Istvanne Rado
Ph.D. h, Stavros Liatis Ph.D. i,
Konstantinos Makrilakis Ph.D. , Luis Moreno Ph.D. j, Katja Wikstro
i €m Ph.D. k, Yannis Manios Ph.D. a,l,*, on
behalf of the Feel4 Diabetes Study Group
a
Department of Nutrition and Dietetics, Harokopio University of Athens, Kallithea, Greece
b
Department of Dietetics, Nutrition and Sport, School of Allied Health, Human Services and Sport, La Trobe University, Melbourne, Australia
c
Neonatal Department, Aretaieio Hospital, Medical School, National and Kapodistrian University of Athens, Athens, Greece
d
Department of Movement and Sports Sciences, Ghent University, Ghent, Belgium
e
Department of Pediatrics, Medical University
Varna, Varna, Bulgaria
f
Clinical Center of Endocrinology and Gerontology, Medical University of Sofia, Sofia, Bulgaria
g
Department of Family and Occupational Medicine, University of Debrecen, Debrecen, Hungary
h
Department of Health Sciences, University of Debrecen, Debrecen, Hungary
i
National and Kapodistrian University of Athens Medical School, First Department of Propaedeutic Medicine, Laiko General Hospital, Athens, Greece
j
Growth, Exercise, Nutrition and Development Research Group, University of Zaragoza, Zaragoza, Spain
k
Population Health Unit, Finnish Institute for Health and Welfare, Helsinki, Finland
l
Institute of Agrifood and Life Sciences, Hellenic Mediterranean University Research Centre, Heraklion, Greece

A R T I C L E I N F O A B S T R A C T

Article History: Objectives: The purpose of this study was to identify lifestyle patterns that are characteristic of overweight
Received 1 February 2022 and obese European adults in the context of educational level.
Received in revised form 22 April 2022 Methods: This was a cross-sectional study of dietary data from 1235 men and 10 343 women. Dietary intake,
Accepted 3 June 2022
educational level, and physical activity were assessed using questionnaires. A principal component analysis
was used to derive lifestyle patterns, and associations with being overweight or obese (OW/OB) and waist
Keywords:
circumference (WC) were explored by applying a multivariate logistic regression.
Diet soft drinks
Results: Overall, 35% of women and 68% of men were OW/OB, of whom 30% and 40%, respectively, had
12 y
Regular soft drinks
Dietary patterns
of education. The principal component analysis derived 2 distinct dietary patterns (healthy vs. unhealthy).
adult The daily intake of fruits (fresh, canned, and juice) and vegetables was found to be associated with lower
Obesity odds of being OW/OB and WC in women only. In contrast, the daily intake of diet soft drinks was associated
Educational level with higher odds of being OW/OB in women and men, but the daily intake of sweets was associated with
higher odds of WC in women only. In both sexes, having >12 y of education was inversely associated with
being OW/OB. No associations were observed for regular soft-drink intake.
Conclusions: A healthy dietary pattern constituting of a daily intake of fruits and vegetables was inversely
associated with being OW/OB in women. Conversely, a diet soft-drink intake was positively associated with
being OW/OB in both sexes, probably because of a reduction in overall energy intake. More studies are rec-
ommended to clarify the effectiveness of diet soft-drink consumption in controlling caloric intake and as a
healthier alternative to regular soft drinks and sweets.
© 2022 Elsevier Inc. All rights reserved.

The Feel4 Diabetes Study has received funding from the EU’s Horizon 2020 research and innovation program (Grant Agreement no. 643708). The content of this article
reflects only the authors’ views, and the EU is not liable for any use that may be made of the information contained herein. The funding body had no role in the design of this
study and collection, analysis, and interpretation of the data, and in writing this manuscript.
*Corresponding author: Prof Yannis Manios, Dept. of Nutrition & Dietetics, School of Health Science & Education, Harokopio University, 70 El. Venizelou Ave, 176 71 Kalli-
thea, Athens, Greece Tel: +30 210 9549156; Fax: +30 2109549141
E-mail address: [email protected] (Y. Manios).

https://doi.org/10.1016/j.nut.2022.111769
0899-9007/© 2022 Elsevier Inc. All rights reserved.
2 K. Reppas et al. / Nutrition 103
104 (2022) 111769

Introduction Finland, Greece, Hungary, Bulgaria, and Spain (National Clinical Trial number, NCT
02393872; https://feel4 diabetes-study.eu/). The Feel4 Diabetes Study was a clus-
ter-randomized design study that consisted of two components: all families and
Overnutrition-related diseases, such as being overweight (OW)
high-diabetes risk families. The objective of this intervention was to develop a
and obese (OB) constitute a major public health threat across the supportive social and physical environment in home and school settings to assist
globe, including the European regions. Worldwide secular trends families in adopting a healthy and active lifestyle. A comprehensive description of
show that the rate of adult overweight or obesity prevalence in 2019 the design, procedures, and implementation of the study have been documented
previously, and only details relevant to this subanalysis are described herein [11].
was the highest in the United States at 73%, followed by the European
Briefly, the sampling of provinces and schools in the present study was random,
Union (EU) at approximately 60%. The prevalence was predominately multistaged, and stratified by socioeconomic indices. Given that screen time is one
in men more than in women, and morbidity was higher in low socio- of the most important energy balance
related behaviors in children [12] and one of
economic groups [1]. In addition, low levels of education were associ- the main behaviors that the Feel4 Diabetes intervention attempted to change, the
ated with a greater risk of being overweight/obese (OW/OB) [2]. sample size calculation was based on screen time and not any other variable. A sam-
ple of 600 families per intervention arm [2] (i.e., 1200 families in total per country)
The accelerating rate of overweight and obese individuals in the was estimated to achieve a statistical power >80% at a two-sided alpha level of 5%
EU is a matter of concern, because apart from premature death [3], to reduce children’s screen time by 0.2 h/d within 8 mo [11,13]. After allowing for a
obesity is linked to a vast range of adverse health outcomes includ- 20% dropout rate, a total sample size of 9000 families was calculated to be required
ing the development of severe COVID 19 symptoms [4], and affects to detect a statistically significant difference between the arms.
The first stage of the intervention involving all families was delivered by
all organ systems, consequentlyleading to a poor quality of life. Obe- teachers within the school setting, and aimed at increasing the consumption of
sity has been estimated to be responsible for >4500 disability- water, fruits, vegetables, breakfast, as well as promoting physical activity and
adjusted life years annually per 100 000 population in the EU [5]. reducing sedentariness. The studied population comprised of 12 041 families
Beyond medical conditions, obesity has a wide social and economic (21 940 parents) and at least one child from each family attending the first three
grades of primary school located in the provinces of six European countries. Base-
effect, including social stigmatization, isolation, depression, anxiety,
line data were collected from April to September 2016. For the purpose of this sub-
health disparities, and employment difficulties, because of discrimi- study, mothers’ and fathers’ data were analyzed only and referred to as women
nation and reduced work attendance [3]. and men throughout the manuscript. Self-administered questionnaires were the
According to the World Obesity Federation, obesity has been only research instrument employed to collect data.
described as a complex relapsing progressive disease driven by an
obesogenic environment in genetically predisposed individuals [3]. Ethical considerations
This study was conducted conforming to the guidelines of the Declaration of
The overconsumption of foods high in calories and nutrient-poor
Helsinki, and all procedures involving human subjects were granted approval by
foods rich in sugars, starch, saturated fat, and salt, combined with the human ethics committees of all of tertiary institutions of the six European
physical inactivity and increased sedentariness, leads to a positive countries before commencement of the study. An extended letter explaining the
energy balance and consequently to an unhealthy accumulation of aims and requirements of the parent study, including a consent form to conduct
fat in adults [3]. Furthermore, the consumption of total, added, and measurements, were provided to the parents and/or guardians of the children in
the schools. Before recruitment into the study, signed consent forms were
simple sugars in the form of sweets, confectionary breakfast cereals, obtained from all legal guardians interested in participating.
sugar-sweetened beverages (SSB), or those naturally occurring in
honey, syrups, and fruit juices, as is typical of a Western-type diet, Instruments
increases the overall energy intake, and may replace the intake of
healthy foods, consequently leading to weight gain in adults [6]. In all centers, at the school site, school teachers and research assistants deliv-
One daily serving of SSB was correlated with a 20% increased likeli- ered self-administered validated standardized questionnaires (translated into
each local language) to parents to collect demographic and dietary information.
hood of developing type 2 diabetes in adults [7]. The World Health
Organization (WHO) recommends that adults and children consume
Demographic characteristics
no more than 10% of their total energy intake (TEI) from added sug- Details on adults’ age was obtained using a four-item Likert scale (<45 y,
ars (equating to 50 g/d for a 2000 kcal/d diet) in the context of a 45
54 y, 55
64 y, and >64 y), which was recoded to the dichotomous variables
healthy diet and up to 5% of TEI for additional health benefits [8]. <45 y and 45 y. Educational level (years of completed schooling) was assessed
Given that foods are not consumed in isolation, dietary pattern by six possible response options (
6 y, 7
9 y, 10
12 y, 13
14 y, 15
16 y, and
>16 y), which were recoded to two categories of
12 y and >12 y, where the
research based on natural eating behavior may be more useful to
threshold of 12 y indicates completion of secondary school. Low socioeconomic
understand the dietary causes of obesity and assist adults in control- status (SES) was defined by
12 y of education attainment [14].
ling weight gain [9]. Ample research has been undertaken on the
intake of total sugars and obesity; however, limited data exist con- Anthropometry
cerning the main food sources of sugars in European populations Adults were measured by rigorously trained researchers, using standardized
[10] and the role of education level in relation to weight status. protocols and equipment in all regions. The physical examination included basic
anthropometric measurements. Subjects were instructed to remove heavy cloth-
Therefore, the purpose of this study was to identify lifestyle
ing and shoes. Weight was recorded to the nearest 0.1 kg using a calibrated SECA
patterns that are characteristic of overweight and obese European digital scale (SECA 813, Hamburg Germany). Height was measured to the nearest
adults in the context of educational level. This study extends our 0.1 cm using a telescopic stadiometer (SECA 213) with subjects standing in an
work on obesity prevention in European populations. The findings erect position without shoes, shoulders relaxed, arms by the side, and head
aligned in the Frankfort plane. Body mass index (BMI), the ratio of weight relative
unveiled in this study are anticipated to provide further evidence-
to height (kg/m2), was used to identify excess body fat in subjects where BMI 25
based knowledge, inspire more research, and be utilized in the kg/m2 signified being overweight and obese according to International Obesity
development of a framework for actions curbing the spread and Task Force thresholds [15].
burden of obesity worldwide. Waist circumference (WC), an indicator of central obesity associated with an
increased risk of type 2 diabetes and cardiovascular disease [16], was assessed
using a non-elastic measuring tape (SECA 201). Measurements were conducted in
Methods subjects wearing minimal clothing, at the end of normal expiration with the arms
relaxed at the side, and the midpoint between the top of the iliac crest and the
Sampling and subjects lower margin of the last rib in the midaxillary line [16]. Then, measurements were
categorized into three categories: <80 cm, 80 to 88 cm, and >88 cm for women
This cross-sectional study was a secondary analysis of baseline data of all fami- and <94 cm, 94 to 102 cm, and >102 cm for men based on the cutoff points advo-
lies participating in the large 2-y pan-European population-based cohort, the Feel4 cated by the WHO for European populations [16].
Diabetes Study (Families across Europe following a healthy Lifestyle 4 Diabetes For analytical purposes, WC categories [3] were merged into two categories:
prevention) undertaken from 2016 to 2018 in six European countries: Belgium, <80 cm versus 80 cm for women and <94 cm versus 94 cm for men [16].
 K. Reppas et al. / Nutrition 103
104 (2022) 111769 3

Similarly, the BMI categories [3] overweight and obese were combined into one

< .001y
< .001
< .001
< .001
< .001
< .001

< .001
< .001

< .001
< .001
group by forming a binary variable: normal weight versus OW/OB.

P*
Lifestyle habits
Regular physical activity was measured with the question “On how many days

12.77 (1360/10 651)

during the last week were you physically active for a total of at least 30 min per

14.76 (1192/8076)

14.83 (1015/6845)

16.40 (1038/6331)
13.93 (519/3726)

12.59 (993/7885)

13.24 (886/6693)
13.20 (859/6506)
day?” The response options were none, 1 d, 2 d, 3 d, 4 d, and 5 d. Data were
recoded into two categories: <3 d and 3 d/wk, with regular physical activity

24.19 § 4.57

26.45 § 3.61
defined as engaging in physical activity 3 d/wk.

Spain

% (n)
Dietary data
Dietary information was derived from adults using a short, validated, 10-item,
food-frequency questionnaire (FFQ) [17], measuring weekly and daily intake of
water, fruit (fresh, canned, juice), vegetables, soft drinks (regular and diet), coffee,

26.19 (2790/10 651)

26.24 (2119/8076)

20.85 (1427/6845)

26.40 (2082/7885)
22.19 (1405/6331)

27.58 (1846/6693)
27.62 (1797/6506)
sweets, salty snacks, or fast food. The FFQ was based on the questionnaire devel-

18.49 (689/3726)
oped for the national type 2 diabetes prevention program in Finland, which has

22.86 § 4.35

27.73 § 4.10
been validated in adults [18].
Typical kitchen measures were used to represent one standard portion size for

Bulgaria
each food item (1 cup, 1/2 cup), as well as commercial units (e.g., 1 slice of pizza, 1/2

% (n)
cup of ice cream). One serving of fruit was considered one medium-sized fruit
item or 1/2 cup, vegetables 1/2 cup, canned or dried fruit 1/2 and 1=4 cup, respectively,
fruit juice 1 cup, regular and diet soft drinks 1 cup, sweets (1 small chocolate bar,
1
/2 cup of ice cream, cookies, or confectionary), and salty snacks/fast food (1 small

15.88 1691/10 651

16.33 (1288/7885)

15.73 (1053/6693)
10.82 (874/8076)

19.14 (713/3726)
13.94 (954/6845)

13.20 (836/6331)

14.37 (935/6506)
hamburger, 1 small bag of chips, 1 slice of pizza). The response options for the fre-
quencies of food intake were <1 time per wk, 1 or 2 times wk, 3 or 4 times per wk,

24.70 § 4.75

27.62 § 4.43
5 or 6 times per wk, 1 or 2 times per d, 3 or 4 times per d, 5 or 6 times per d, and

Hungary
>6 times per d. Based on WHO healthy eating guidelines for adult populations
[19] and previous research [20], food intake frequencies were recoded into two

% (n)
categories: daily and less than daily for all food items. Collectively, the complete

Country
dietary data were available for 11 578 adults (men: n = 1235; women: n = 10 343).

17.93 (1910/10 651)

Statistical analysis

15.12 (1221/8076)

20.25 (1386/6845)

16.31 (1286/7885)

21.17 (1417/6693)
20.37 (1325/6506)
21.26 (792/3726)

12.81 (811/6331)
All statistical computations were performed using SPSS, version 27 (IBM Corp.,

24.49 § 4.71

27.73 § 4.05
Chicago, IL). Continuous variables were checked for normality using the Kolmogor-
ov
Smirnov test and group differences applying non-parametric the Pearson’s x2
Greece

and Kruskal
Wallis tests. Continuous data are presented as means and standard
% (n)

deviations, and categorical data as frequencies and percentages. To derive lifestyle

patterns, a categorical principal component analysis was conducted using the 11-
item questionnaire that measured frequency of food intake and physical activity
11.83 (1260/10 651)

level in adults. The suitability of the PCA was evaluated by inspection of the corre-
16.57 (1338/8076)

18.58 (1176/6331)
12.05 (449/3726)
12.46 (853/6845)

11.53 (909/7885)

10.67 (694/6506)
lation matrix, which showed that none of the variables had correlation coefficients

9.62 (644/6693)
exceeding the threshold of 0.32 [21]. Based on this assumption, the PCA was con-
24.90 § 4.60

27.14 § 3.66
ducted applying the orthogonal VARIMAX rotation with the Kaiser normalization.
After a series of data reductions commencing from a five-factor analysis, a two-
Finland
Descriptive characteristics of study participants for the total sample and by country of residence.

component solution met the interpretability criterion [22] and was retained. As
% (n)

proposed by Kline [23], the loading significance was set at 0.3.

The degree of internal consistency for each dimension and in total, as well as the
reliability of the questionnaire, were measured with Cronbach’s alpha coefficient,
15.40 (1640/10 651)

where a value of 0.7 is considered acceptable [24]. Then, the logistic regression
16.49 (1332/8076)

17.68 (1210/6845)

16.83 (1327/7885)
16.82 (1065/6331)

12.66 (847/6,693)
15.14 (564/3726)

13.77 (896/6506)

model and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to esti-
mate the associations between dietary patterns and OW/OB. In total, four multivariate
24.12 § 4.12

25.77 § 3.54

analyses were performed, one for each dimension, and using women’s and men’s data
Belgium

separately. BMI categories were entered as the dependent variable and components of
the two dimensions as independent variables. Normal weight and less than daily fre-
% (n)

quency of food intake were assigned as the reference groups. The regression model
*P- value for country differences estimated using Pearson's X2 test.

was adjusted for age group, educational level, and regular physical activity [25].
Exploratory to this study was the investigation of dietary patterns and central
90.36 (10 651/11 787)

obesity (represented by WC), applying the regression model adjusted for the same
71.29 (8076/11 328)

34.41 (3726/10 828)

59.27 (6845/11 548)

77.66 (7885/10 153)

confounding factors. The predictive accuracy of the regression model was investi-
64.12 (6331/9873)

68.46 (6693/9777)
65.55 (6506/9926)

gated with inspection of the classification table. A two-tailed P value was consid-
ered significant at the 5% level.
24.05 § 4.57

27.16 § 4.02

BMI, body mass index; WC, waist circumference.

Total % (N)

In bold, statistically significant P values at 5%.

Results
Kruskal-Wallis non-parametric test

Of the 21 940 adults recruited at baseline, 54% were women

(n = 11 787) and 46% were men (n = 10 153). In total, the majority of
Tertiary Education (>12 y)

Tertiary education (>12 y)

adults (90%) were age <45 y, and more than half (60%) had com-
BMI, kg/m2, mean § SD

BMI, kg/m2, mean § SD

pleted tertiary education (>12 y). With regard to weight status,

OW/OB (25 kg/m2)

OW/OB (25 kg/m2)

approximately 35% of women and 68% of men were categorized as

OW/OB, and 60% of women had WC 80 cm and 65% of men
Characteristics

WC 80 cm

WC 94 cm

94 cm. Approximately 30% of women and 40% of men who were
Age <45 y

Age <45 y
Women

OW/OB were of low SES (women: 33.9% [n = 1192 of 3509];

Table 1

Men

P < 0.001; men: 38.4% [n = 2392 of 6236]; P < 0.001; data not
shown).
y
4 K. Reppas et al. / Nutrition 103
104 (2022) 111769

The descriptive characteristics of the study participants for the obesity (Table 3). More specifically, in the crude regression analysis, a
total sample and by country of residence are reported in Table 1. daily intake of fresh fruit and vegetables was associated with lower
There were significant country variations in the prevalence rate of odds of women being OW/OB compared with those consuming these
OW/OB and central obesity, with considerably higher rates foods less than daily (fresh fruit: OR: 0.84; 95% CI, 0.76
0.93; vegeta-
observed in Greek women (OW/OB: 21.3%; WC 80 cm: 20.2%; bles: OR: 0.81; 95% CI, 0.73
0.90). After adjusting for regular physical
P < 0.001) and Bulgarian men (OW/OB: 27.6%; WC 94 cm: 27.6%; activity, age group, and educational level, significance remained for
P < 0.001). Notably, almost half of Greek women and Bulgarian fresh fruit (OR: 0.86; 95% CI, 0.77
0.96) and vegetables (OR: 0.86; 95%
men who were OW/OB belonged to the low SES group (Greek CI, 0.77
0.96), but was borderline for the daily intake of fruit juice
women: 46.7% [3n = 64 of 780], P < 0.001; Bulgarian men: 45.1% (OR: 0.89; 95% CI, 0.78
1.00).
[n = 804 of 1782], P < 0.001; data not shown). Interestingly, women engaging in regular physical activity and
With respect to the dietary habits of adults, apart for daily those with a high educational level (>12 y) were associated with
intake of water, canned fruit, and sweets, marked differences were lower odds of being OW/OB (i.e., prophylactic effect; regular physi-
observed between normal weight and OW/OB women for daily cal activity: OR: 0.80; 95% CI, 0.72
0.88; high education level: OR:
intake of fresh fruit (P < 0.001), fruit juice (P = 0.033), regular soft 0.72; 95% CI, 0.64
0.80), but being age >45 y was positively associ-
drinks (P = 0.004), diet soft drinks (P < 0.001), coffee (P = 0.028), ated with being OW/OB (OR: 1.58; 95% CI, 1.35
1.84). with regard
vegetables (P < 0.001), and salty snacks/fast food (P = 0.02; Supple- to central obesity, the same trend was observed for associations
mentary Table 1). In contrast, no differences were observed between fresh fruit and vegetables in the unadjusted analysis
between normal-weight versus OW/OB men, except for daily (fresh fruit: OR: 0.89; 95% CI, 0.80
0.98; vegetables: OR: 0.87; 95%
intake of diet soft drinks, which was considerably higher in the lat- CI, 0.79
0.97), although associations for the latter became mar-
ter (7.0% vs. 15.0%; P < 0.001; Supplementary Table 2). When com- ginal in the adjusted model (vegetables: OR: 0.91; 95% CI,
paring dietary intake between OW/OB women and men, no 0.82
1.01).
differences were observed (data not shown; Supplementary Fig. 1). Nevertheless, negative associations were observed for canned
fruit, fruit juice, and WC (canned fruit: adjusted OR [aOR]: 80; 95%
Principal component analysis CI, 0.69
0.93); fruit juice: aOR: 0.81; 95% CI, 0.72
0.90). Women
aged 45 y were associated with higher odds of central obesity
The PCA yielded two dimensions that had eigenvalues >1. Of the (aOR: 1.53; 95% CI, 1.31
1.79), but regular physical activity (aOR:
11 components, coffee (0.137) and regular exercise (0.197) had load- 0.74; 95% CI, 0.67
0.81) and educational level >12 y (aOR: 0.88;
ings <0.3; therefore, a second analysis was conducted using the VARI- 95% CI, 0.80
0.98) with lower odds.
MAX rotation excluding these two components. Two dimensions were
derived indicating two distinct dietary patterns: dimension 1, a healthy Associations between unhealthy dietary pattern in women versus
dietary pattern consisting of daily intake of water, fruit (fresh, canned, OW/OB and WC
juice), and vegetables versus dimension 2, unhealthy dietary pattern of With respect to unhealthy dietary patterns, the daily consump-
regular soft drinks, diet soft drinks, sweets, and salty snacks/fast food. tion of diet soft drinks was associated with higher odds of being
The value of Cronbach’s alpha of 0.793 indicated good reliability of the OW/OB in women (OR: 1.65; 95% CI, 1.42
1.91), but daily sweet
model [24]. The percentage of variance accounted for by dimensions 1 intake was associated with lower odds (OR: 0.87; 95% CI,
and 2 were 18.9% and 18.8%, respectively, and 37.7% for the total 0.79
0.96; Table 3). After adjusting for regular physical activity,
model, which increased from 31.0% in the first rotation. A table of the age group, and educational level, the results remained unchanged
rotated loadings matrix with unsuppressed coefficients is presented in (diet soft drinks: OR: 1.67; 95% CI, 1.44
1.95; sweets: OR: 0.88;
Table 2. A visual representation of the biplot component loadings 95% CI, 0.79
0.98). No significant results were observed for salty
derived from the VARIMAX rotation is shown in Figure 1. snacks/fast food or regular soft drinks.
With respect to central obesity, comparable findings were
Multivariate logistic regression models found. In the adjusted analysis, a positive association was observed
between daily intake of diet soft drinks and WC 80 cm (OR: 1.65;
95% CI, 1.41
1.92). However, unlike associations for OW/OB, the
Associations between healthy dietary pattern in women versus OW/ reverse was noted for the daily intake of sweets, salty snacks/fast
OB and WC food, and central obesity. The daily intake of sweets was associated
The logistic regression showed that women adhering to a healthy with higher odds of WC 80 cm (OR: 1.15; 95% CI, 1.05
1.27) and
dietary pattern had a prophylactic effect on OW/OB, as well as central salty snacks/fast food with lower odds (OR: 0.82; 95% CI,
Table 2 0.68
0.99).
Rotated component loadings derived using VARIMAX rotation Consistent with healthy dietary patterns, regular physical activ-
ity and high educational level were associated with lower odds of
Rotated component loadings*
OW/OB and central obesity in women (regular physical activity
Dimension OW/OB: OR: 0.79; 95% CI, 0.71
0.87; WC 80 cm: OR: 0.71; 95%
Components/variables (daily intake) 1 2 CI, 0.64
0.77; high educational level: OW/OB: OR: 0.67; 95% CI,
0.60
0.75; WC 80 cm: OR: 0.83; 95% CI, 0.74
0.92), whereas age
Fresh fruit 0.760y
0.033
Vegetables 0.657y 0.046
>45 y was associated with higher odds (age group >45 y OW/OB:
Water 0.624y
0.359 OR: 1.63; 95% CI, 1.39
1.91; WC 80 cm: OR: 1.57; 95% CI,
Canned fruit 0.407y 0.330 1.33
1.84). An inspection of the classification table showed that
Fruit juice 0.319y 0.273 there was 67% reliability/accuracy of the adjusted healthy and
Salty snacks/fast food 0.032 0.668y
unhealthy dietary models in explaining OW/OB in women.
Regular soft drinks
0.126 0.667y
Sweets 0.126 0.544y
Diet soft drinks
0.008 0.435y Associations between dietary patterns in men and WC
**Rotation method: VARIMAX with Kaiser normalization. In contrast with women, in both the crude and adjusted multi-
y
In bold, loadings 0.32. variate analyses, no associations were found for adherence to a
 K. Reppas et al. / Nutrition 103
104 (2022) 111769 5

Fig. 1. Biplot component loadings displaying healthy and unhealthy dietary patterns.

healthy dietary pattern in men and OW/OB or central obesity model had 67.5% reliability in predicting OW/OB in men. When
(Table 4). However, an unhealthy dietary pattern comprising of a repeating the regression model to explore associations between
daily intake of diet soft drinks promoted OW/OB in men (diet soft the unhealthy dietary pattern and WC 94 cm, null associations
drinks: crude: OR: 2.25; 95% CI, 1.40
3.63; aOR: 2.22; 95% CI, were found.
1.36
3.61). In agreement with women’s data, men with >12 y of
education were inversely associated with OW/OB (OR: 0.60; 95% Discussion
CI, 0.44
0.82). No relationship was observed for age group
(adjusted P = 0.11) or regular physical activity (adjusted P = 0.42). The findings of the present study confirm the challenge of over-
An inspection of the classification table revealed that the unhealthy nutrition and obesity in many EU member states, indicating the

Table 3
Associations between healthy and unhealthy dietary patterns in women, OW/OB, and central obesity from the multivariate logistic regression analysis

Women (n = 10 343)

Healthy dietary pattern

OW/OB Central obesity

(BMI 25 kg/m2) (WC 80 cm)

Model 1 Model 2 Model 1 Model 2

Food items (daily intake) OR 95% CI P* OR 95% CI Adjusted Py OR 95% CI P* OR 95% CI Adjusted Py

Water 1.1 0.83
1.46 .486 1.2 0.90
1.61 .219 1.14 0.88
1.48 .321 1.18 0.91
1.54 .217
Fresh fruit 0.84 0.76
0.93 .001 0.86 0.77
0.96 .006 0.89 0.80
0.98 .018 0.9 0.81
0.99 .041
Canned fruit 1 0.85
1.18 .973 0.98 0.83
1.16 .823 0.78 0.68
0.90 .001 0.8 0.69
0.93 .004
Fruit juice 0.91 0.81
1.03 .151 0.89 0.78
1.00 .059 0.81 0.72
0.90 < .001 0.81 0.72
0.90 < .001
Vegetables 0.81 0.73
0.90 < .001 0.86 0.77
0.96 .008 0.87 0.79
0.97 .012 0.91 0.82
1.01 0.086

Unhealthy dietary pattern

OW/OB Central obesity

(BMI 25 kg/m2) (WC 80 cm)

Model 1 Model 2 Model 1 Model 2

Food items (daily intake) OR 95% CI P* OR 95% CI Adjusted Py OR 95% CI P* OR 95% CI Adjusted Py

Regular soft drinks 1.12 0.97
1.29 .111 1.08 0.93
1.25 .286 0.98 0.860
1.13 .817 0.96 0.84
1.11 .612
Diet soft 1.65 1.42
1.91 < .001 1.67 1.44
1.95 < .001 1.59 1.37
1.85 < .001 1.65 1.41
1.92 < .001
Sweets 0.87 0.79
0.96 .008 0.88 0.79
0.98 .017 1.14 1.04
1.25 .006 1.15 1.05
1.27) .004
Salty snacks/fast food 1.17 0.97
1.42 .091 1.13 0.93
1.37 .225 0.84 0.71
1.01 .061 0.82 0.68
0.99 .038
BMI, body mass index; CI, confidence interval; OR, odds ratio; OW/OB, overweight/obesity; WC, waist circumference.
In bold, text statistically significant P values at 5%. Food item variables derived from food-frequency questionnaire: 0 is < daily and 1 is daily. Reference: first category < daily
is 0. Dependent variable 1, women’s BMI category normal weight (0) versus OW/OB (1). Dependent variable 2, WC <80 cm (0) versus WC 80 cm (1). Independent variable 1,
healthy dietary pattern components: daily intake of fresh fruit, canned fruit, fruit juice, and vegetables. Independent variable 2, unhealthy dietary pattern components: daily
intake of regular and diet soft drinks, sweets, and salty snacks/fast food.
**Model 1 P value estimated from crude logistic regression model.
y
Model 2 P value adjusted for regular physical activity, women’s age group, and education level.
6 K. Reppas et al. / Nutrition 103
104 (2022) 111769

Table 4
Associations between healthy and unhealthy dietary patterns, OW/OB, and central obesity in men from the multivariate logistic regression analysis

Men (n = 1235)

Healthy dietary pattern

OW/OB Central obesity

(BMI 25 kg/m2) (WC 94 cm)

Model 1 Model 2 Model 1 Model 2

Food intake (daily) OR 95% CI P* OR 95% CI Adjusted Py OR 95% CI P* OR 95% CI Adjusted Py

Water 1.41 0.65
3.08 .385 1.57 0.71
3.50 .266 1.22 0.58
2.59 .597 1.22 0.56
2.70 .62
Fresh fruit 1.05 0.80
1.39 .708 1.06 0.80
1.41 .688 0.9 0.69
1.18 .46 0.92 0.70
1.22 .581
Canned fruit 0.85 0.52
1.38 .508 0.78 0.47
1.29 .335 0.78 0.49
1.24 .298 0.72 0.45
1.18 .193
Fruit juice 1.01 0.73
1.41) .929 0.99 0.71
1.40 .973 0.83 0.61
1.14 .247 0.8 0.58
1.11 .183
Vegetables 0.88 0.66
1.16) .354 0.93 0.69
1.24 .62 1.01 0.77
1.32 .93 1.08 0.81
1.43 .594

Unhealthy dietary pattern

OW/OB Central obesity

(BMI 25 kg/m2) (WC 94 cm)

Model 1 Model 2 Model 1 Model 2

y
Food intake (daily) OR 95% CI P* OR 95% CI Adjusted P OR 95% CI P* OR 95% CI Adjusted Py

Regular soft drinks 0.87 0.62
1.21 .407 0.8 0.56
1.14 .222 0.94 0.68
1.29 .69 0.92 0.65
1.29 .612
Diet soft drinks 2.25 1.40
3.63 .001 2.22 1.36
3.61 .001 1.41 0.94
2.12 .098 1.4 0.92
2.13 .117
Sweets 0.98 0.72
1.34 .902 0.97 0.70
1.33 .835 1.08 0.79
1.46 .633 1.1 0.80
1.50 .562
Salty snacks/fast food 1.01 0.59
1.73 .961 1 0.57
1.74 .999 0.82 0.50
1.34 .423 0.83 0.49
1.38 .468
BMI, body mass index; CI, confidence interval; OR, odds ratio; OW/OB, overweight/obesity; WC, waist circumference.
In bold, statistically significant P values at 5%. Dependent variable 1, men’s BMI category (normal weight = 0 vs. overweight/obese = 1). Dependent variable 2, men’s WC groups (0
is <94 cm and 1 is 94 cm). Independent variable 1, healthy dietary pattern components: Daily intake of water, fruit (fresh, canned, juice), and vegetables. Independent variable
2, unhealthy dietary pattern components: Daily intake of regular and diet soft drinks, sweets, and salty snacks/fast food. Reference: First group < daily intake is 0.
**Model 1 P value estimated from crude logistic regression analysis.
y
Model 2 P value adjusted for men’s age group, regular physical activity, and education level.

complexity and severity of this condition. From a large sample of fast food, and diet soft drinks and OW/OB and WC in women, but
adults residing in six European countries, we observed that 35% of only for diet soft drinks and OW/OB in men after adjusting for age
women and 68% men were classified as having BMI 25 kg/m2, group, regular physical activity, and education level. By the size of
and approximately 60% of men or women presented central obe- the OR, there was a stronger association of daily diet soft drink
sity based on WHO standards. Country differences in the distribu- intake and OW/OB in men (aOR: 2.22) than in women (aOR: 1.67).
tion of OW/OB and central obesity revealed that both indices were This study supports the notion that diet soft-drink intake is related
the highest in Greek women and Bulgarian men. In fact, the per- to OW/OB in adults and of high relevance, especially when obesity
centage of OW/OB in Greek and Bulgarian populations ranged prevalence continues to accelerate and artificially sweetened bev-
from 21% to 27% of men and 18% to 21% of women, with almost erage (ASB) intake is high. This is not surprising and in accordance
half of those OW/OB belonging to the low SES group as defined by with the literature database [27].
educational level. Nonetheless, these results need to be interpreted with caution.
Our findings are in line with the latest data from 27 countries Although this observation is consistent with those of previous
collated by the European Commission of Statistics (Eurostat) in studies suggesting an association between ASBs and adult obesity
2019, reporting that 53% of adults age >18 y (60% men; 46% [27], diet soft drinks may be associated with weight gain by reverse
women) residing in the EU were considered OW/OB based on BMI causation. Individuals who are OW/OB may replace SSB with ASB
[26]. Furthermore, the problem of OW/OB was recorded in >50% of in an effort to balance the high-caloric density of an unhealthy die-
populations bordering the Balkan peninsula and Mediterranean tary pattern. The presence of this confounding factor may overesti-
basin [26]. Astonishingly, 67% of men in Greece were OW/OB and mate the strength of the association between daily diet soft drink
64% in Bulgaria compared with 49% and 46% of women, respec- intake and OW/OB in adults. Even so, our findings are important
tively [26]. Comparable with our data, convincing evidence con- given that diet soft-drink consumption is gaining popularity.
firms a decreasing gradient between OW/OB and SES. Up-to-date Clearly, more cohort studies are needed to validate our results and
statistics from Eurostat illustrate that the rate of OW/OB in adults examine the effectiveness of diet soft-drink consumption in con-
was 59% and 20% among those of low education level, 54% and 17% trolling caloric intake and as a healthier alternative to regular soft
for those of medium education level, and 44% and 11% for adults of drinks or sweets in dietary interventions battling adult obesity.
high education level [26]. This portrays the high diversity of health The adverse health effects of excessive sugar intake are well-rec-
inequalities within the EU and the gap in knowledge of the drivers ognized [8]. Nevertheless, whether ASBs are detrimental to health is
that increase obesity in vulnerable subgroups of low educational debatable [27,28]. In a meta-analysis of pooled data from 39 pro-
level. Indeed, SES inequalities should be accounted for by policy- spective cohort studies, Qin et al. demonstrated that, for every
makers when developing population-based obesity prevention increase in SSB and ASBs by 250 mL per d, there was a 12% and 21%
strategies. increase, respectively, in risk for obesity, 19% and 15% increase for
The second outcome of the present study was that the regres- type 2 diabetes, 10% and 8% increase for hypertension, as well as 4%
sion analysis revealed positive associations between an unhealthy and 6% increase for all-cause mortality [27]. In a recent cohort study
dietary pattern consisting of a daily intake of sweets, salty snacks/ by Chazelas et al. including 104 760 adults, high consumers of both
 K. Reppas et al. / Nutrition 103
104 (2022) 111769 7

SSBs and ASBs had 1.2 and 1.32 times a greater risk of experiencing findings of a recent review documenting that moderate intakes of
a cardiovascular event than non-consumers [28]. Surprisingly, less fruit juice (75
224 mL/d) were not associated with an increased
frequent consumers of either SSBs or ASBs had 1.05 and 1.15 times a risk of obesity [40]. Fruit juice is considered a healthier alternative
greater risk than non-consumers. to SSB and ASBs. From a practical point of view, our observation sug-
There are several biologic mechanisms that can explain the det- gests that even processed foods, provided that they are consumed in
rimental effect of ASB intake in relation to lipogenesis involving moderation within the context of a healthy diet, might prevent cen-
appetite dysregulation, hormonal control, and gut microbiome tral obesity and warrants consideration in future studies.
perturbation [29]. The sweet taste of artificial sweeteners may A likely explanation for null associations between a daily intake
elicit metabolic changes activated by sweet receptors in the oral of FV and OW/OB or central obesity in men might be related to low
cavity and tissues within the intestine, pancreatic b cells, and the power, because the frequency of men who provided dietary details
brain, as well as alterations in the gut microbiota induced by glu- was considerably less than that of women (n = 1235 vs. 10 343).
cose intolerance [29]. Alternatively, compared with solid foods, the Alternatively, misclassification due to misreporting errors of FV
consumption of beverages has a lower pronounced effect on the would bias the results toward null values. Furthermore, compared
level of satiety/fullness with a more rapid recurring feeling of hun- with OW/OB women, a low habitual intake of FV has been reported
ger; thus, resulting in increased energy intake [30]. On the other in overweight men [41]. In our study, approximately 60% of OW/OB
hand, diet soft drinks are advertised as safe, healthy alternatives to men consumed daily FV; thus, suggesting that sex-specific strate-
SSBs by retaining sweetness without increasing caloric intake, and gies promoting the intake of healthy foods may be necessary to
recommended by health professionals for weight loss [31]. This reduce OW/OB, especially in men. Our findings correspond with
brings attention to the need for future public health campaigns contemporary data from Eurostat reporting that in 2019, only
focusing on reducing SSBs/ASBs and their substitution with health- 51.7% of men from 27 EU countries consumed one to four portions
ier drinking options, such as water, as a positive step toward curb- of FV daily, and an appalling 39% consumed none at all [42]. Fur-
ing the obesity epidemic in the EU. ther work is needed in the EU to promote the consumption of FV
In our study, regarding the negative associations observed to 400 g/d (five portions/d) as endorsed by the WHO for the pre-
between daily intake of sweets and OW/OB, as well as salty snacks/ vention of chronic disease [43].
fast-food intake and central obesity in women, feasibly underre- To summarize, based on the observations of the present study,
porting of energy-dense foods is common practice in subjects with one could speculate that a more pronounced change of OW/OB and
excess weight because of social desirability [32]. Nonetheless, we WC could be obtained when foods preventing body fat gain in a diet
found that a daily intake of sweets was positively associated with displace those promoting abdominal fat and weight gain. Only then
WC 80 cm, which is credible and supported by strong evidence will strategies promoting FV be successful if guidelines are provided
from ad libitum trials reporting positive associations between a for the decreased consumption of foods that enhance weight gain.
high intake of sugars and obesity in adults [33]. Sugar or sucrose, Of note, in the adjusted regression models, women aged >45 y
which is added in the manufacture of sweets and confectionary, is showed strengthened associations for OW/OB, as well as central
a disaccharide of glucose and fructose bound by an a-1,2-glyco- obesity, for both dietary patterns whereas regular physical activity
sidic link, and activates an increased secretion of gastrointestinal and high education attainment (>12 y of schooling) showed a lesser
glucose-dependent insulinotropic peptide independent of its likelihood of being OW/OB (protective effect). Unexpectedly, only
energy content, which in turn increases appetite and overcon- high educational level was found to be inversely associated with
sumption of calories and consequently increases body weight and OW/OB for men, suggesting that young women are less likely to suf-
obesity [34]. fer from obesity than middle-aged women along with a prophylactic
Another significant outcome of this study was that a healthy effect of high education attainment and regular physical activity on
dietary pattern of a daily intake of fresh fruits and vegetables (FV) the probability of being OW/OB [25].
was inversely associated with OW/OB and WC in women only, These observations are consistent with data from the 2019
although vegetable intake was marginally significant after adjust- European Health Interview Survey reporting variations in over-
ing for women’s age group, regular physical activity, and educa- weight status by sex, age, and education level [44]. One could spec-
tional level. A pooled analysis of 17 prospective cohort studies ulate that low income and education level, lack of nutrition
including 563 277 participants showed that intake of FV was asso- knowledge, and health illiteracy render cheap, unhealthy, tasty
ciated with modest reductions in body weight and WC, along with convenient foods an attractive alternative to expensive healthy
a 17% reduced risk of adiposity [35]. The mechanisms by which FV foods, such as FV [45,46]. This is important because a regular intake
promote and assist in weight loss and the maintenance of a healthy of FV is associated with a reduced risk of obesity, cardiovascular
weight are attributed to the very low energy density, high soluble disease, diabetes, and various cancers attributed to the high con-
fiber, and water content that produce a satiating effect, improved tent of vitamins, minerals, antioxidants, and bioactive compounds
glycemic control, and consequently reduced consumption of that have a myriad of health benefits [43].
energy-dense foods; thus, decreasing TEI [36,37]. In reference to our finding that regular physical activity
In addition, FV are abundant in vitamins, minerals, and phyto- (30 min of physical activity at least 3 times/wk) was related to
chemicals that have antioxidant and antiinflammatory properties lower odds of OW/OB and central obesity in women, weightloss
and act synergistically against obesity-induced oxidative stress therapies recommend lifestyle interventions that include energy-
and inflammation [38]. Emerging evidence linking the gut micro- restricted diets and participation in regular physical activity [47].
biome to obesity indicates that high-fiber diets are associated with Regular physical activity is well-recognized to be associated with
a healthier microbiome of rich biodiversity and stimulate a higher an abundance of beneficial health effects, such as increased aerobic
production of short-chain fatty acids, butyrate, and propionate, capacity and muscle strength, reduced risk of chronic diseases (e.
which exert antiobesogenic properties and are associated with a g., type 2 diabetes, hypertension, dyslipidemia, and cardiovascular
lower risk of obesity [39]. disease), along with stress levels and improved mood [48]. Even
Intriguingly, we found that a daily intake of canned fruit (1/2 cup) more important than metabolic health, physical activity causes
and fruit juice (1 cup) were related to approximately 20% lower changes in body fat composition, including a reduction in abdomi-
odds of central obesity in women, which is supported by the nal visceral adiposity and total body fatness [48]. Current
8 K. Reppas et al. / Nutrition 103
104 (2022) 111769

guidelines for the treatment of OW/OB advocate >250 min/wk of University (Greece): Yannis Manios, Kalliopi Karatzi, Odysseas
physical activity to achieve clinically significant weight loss and Androutsos, George Moschonis, Spyridon Kanellakis, Christina
from 150 to 250 min/wk to prevent weight gain [48]. Mavrogianni, Konstantina Tsoutsoulopoulou, Christina Katsarou,
Given that a healthy diet, combined with increased physical Eva Karaglani, Irini Qira, Efstathios Skoufas, Konstantina Maragko-
activity, is the prescription for weight loss and maintenance of a poulou, Antigone Tsiafitsa, Irini Sotiropoulou, Michalis Tsolakos,
healthy weight, our findings amplify the urgent need to reduce Effie Argyri, Mary Nikolaou, Eleni-Anna Vampouli, Christina Fili-
health and food inequalities in vulnerable groups residing in the ppou, Kyriaki Apergi, Amalia Filippou, Gatsiou Katerina, and Efstra-
EU by making healthy food sources affordable, available, and acces- tios Dimitriadis; Finnish Institute for Health and Welfare (Finland):
sible to all, as well as the development of safe recreational areas for Jaana Lindstro € m, Tiina Laatikainen, Katja Wikstro € m, Jemina Kivela€,
physical activity in low SES suburbs. Pa€ivi Valve, Esko Leva €lahti, Eeva Virtanen, Tiina Pennanen, Seija
Olli, and Karoliina Nelimarkka; Ghent University (Belgium),
Strengths and limitations Department of Movement and Sports Sciences: Greet Cardon,
Vicky Van Stappen, and Nele Huys; Ghent University (Belgium),
To the best of our knowledge, this is the first comprehensive Department of Public Health: Lieven Annemans and Ruben Wil-
study of sex-specific associations of socio-lifestyle factors with lems; Ghent University (Belgium), Department of Endocrinology
overweight and obesity tested in a European cohort study. A major and Metabolic Diseases: Samyah Shadid; Technische Universita €t
strength of the present study was the large sample size, along with Dresden (Germany): Peter Schwarz and Patrick Timpel; University
the random and multistage sampling of schools. Our novel study of Athens (Greece): Konstantinos Makrilakis, Stavros Liatis, George
identified a range of modifiable dietary risk factors that could aid Dafoulas, Christina-Paulina Lambrinou, and Angeliki Giannopou-
in the prevention of adult OW/OB. Most studies have explored the lou; International Diabetes Federation European Region (Belgium):
effect of overall FV group intake and sugar-sweetened beverages in Winne Ko and Ernest Karuranga; Universidad De Zaragoza (Spain):
relation to adult adiposity, as measured by BMI or weight change Luis Moreno, Fernando Civeira, Gloria Bueno, Pilar De Miguel-
[33,35,49], and our study adds new data to the limited database. In Etayo, Esther M’ Gonzalez-Gil, María L. Miguel-Berges, Natalia
addition, body composition was measured objectively using two Gime
nez-Legarre; Paloma Flores-Barrantes, Alelí M. Ayala-Marín,
tools: BMI and WC. The latter is considered a reliable indicator of Miguel Seral-Corte
s, Lucia Baila-Rueda, Ana Cenarro, Estíbaliz Jar-
visceral fat [16], is easily measured in subjects, can complement auta, and Rocío Mateo-Gallego; Medical University of Varna (Bul-
BMI estimates, and provides a more complete picture of the magni- garia): Violeta Iotova, Tsvetalina Tankova, Natalia Usheva, Kaloyan
tude/severity of adult OW/OB in six EU member states. Tsochev, Nevena Chakarova, Sonya Galcheva, Rumyana Dimova,
Our study should be viewed in light of some limitations. The Yana Bocheva, Zhaneta Radkova, Vanya Marinova, Yuliya Bazdar-
cross-sectional nature of the study design hinders the establish- ska, and Tanya Stefanova; University of Debrecen (Hungary): Imre
ment of a causal link between the specific food patterns and adult Rurik, Timea Ungvari, Zolta
n Jancso
, Anna Na
na

si, L
aszlo


body fatness. The limited number of foods assessed in our FFQ Kolozsva
Bíro

ri, Csilla Sem
anova, Eva
, Emese Antal, and Sa
ndorne

could have led to missed associations between other foods and Rado
; and Extensive Life Oy (Finland): Remberto Martinez and
adult obesity. For example, regular alcohol intake has been linked Marcos Tong.
to a higher BMI and visceral obesity in men [50]. We cannot
exclude reporting bias because habitual food intake was self- Supplementary materials
reported by adults. Nevertheless, questionnaires were identical in
all countries and tested for accuracy and validity before baseline Supplementary material associated with this article can be
data collection [17]. One more drawback is that soft-drink con- found in the online version at doi:10.1016/j.nut.2022.111769.
sumption was surveyed in terms of frequency rather than volume,
and TEI of subjects was not quantifiable because of data collection
References
constraints. Therefore, whether diet soft-drink intake contributes
to adult OW/OB warrants further investigation. Replication in
[1] Organisation for Economic Cooperation and Development. Overweight and
other large prospective cohorts and mechanistic investigations are obesity among adults. Health at a Glance: OECD indicator. Available at:
needed. https://www.oecd-ilibrary.org/sites/0f705cf8-en/index.html?itemId=/content/
component/0f705cf8-en. Accessed December 1, 2021.
[2] Kim TJ, Roesler NM, von dem Knesebeck O. Causation or selection
examining
Conclusions the relation between education and overweight/obesity in prospective obser-
vational studies: a meta-analysis. Obes Rev 2017;18:660–72.
The findings of this study suggest that a healthy dietary pattern [3] Bray GA, Kim KK, Wilding JPH, World Obesity Federation. Obesity: a chronic
relapsing progressive disease process. A position statement of the World Obe-
constituting of a daily intake of FV was inversely associated with sity Federation. Obes Rev 2017;18:715–23.
OW/OB in women. Conversely, diet soft-drink intake was posi- [4] Katz MH. Regardless of age, obesity and hypertension increase risks with
tively associated with OW/OB in both sexes, probably in an effort COVID-19. JAMA Intern Med 2021;181:381.
[5] Organisation for Economic Cooperation and Development. The heavy burden
to reduce overall energy intake rather than the inverse. More stud- of obesity. Paris, France: OECD Publishing; 2019.
ies are recommended to clarify the effectiveness of diet soft-drink [6] Luger M, Lafontan M, Bes-Rastrollo M, Winzer E, Yumuk V, Farpour-Lambert N.
consumption in controlling caloric intake and as a healthier alter- Sugar-sweetened beverages and weight gain in children and adults: a system-
atic review from 2013 to 2015 and a comparison with previous studies. Obes
native to regular soft drinks and sweets in the battle against adult
Facts 2017;10:674–93.
obesity. €fvenborg JE, Andersson T, Carlsson PO, Dorkhan M, Groop L, Martinell M,
[7] Lo
et al. Sweetened beverage intake and risk of latent autoimmune diabetes in
adults (LADA) and type 2 diabetes. Eur J Endocrinol 2016;175:605–14.
Acknowledgments
[8] World Health Organization. Guideline: sugar intake for adults and children.
Available at: https://www.who.int/publications/i/item/9789241549028.
The authors thank the members of the Feel4 Diabetes Study Accessed December 1, 2021.
Group. Coordinator: Yannis Manios; Steering Committee: Yannis [9] Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology.
Curr Opin Lipidol 2002;13:3–9.
€m, Peter Schwarz, Konstanti-
Manios, Greet Cardon, Jaana Lindstro [10] Newens KJ, Walton J. A review of sugar consumption from nationally represen-
nos Makrilakis, Lieven Annemans, and Winne Ko; Harokopio tative dietary surveys across the world. J Hum Nutr Diet 2016;29:225–40.
 K. Reppas et al. / Nutrition 103
104 (2022) 111769 9

[11] Manios Y, Androutsos O, Lambrinou CP, Cardon G, Lindstrom J, Annemans L, [31] Johnson RK, Lichtenstein AH, Anderson CAM, Carson JA, Despre
s JP, Hu FB,
et al. A school- and community-based intervention to promote healthy life- et al. Low-calorie sweetened beverages and cardiometabolic health: a science
style and prevent type 2 diabetes in vulnerable families across Europe: design advisory from the. American Heart Association. Circulation 2018;138:e126–
and implementation of the Feel4 Diabetes Study. Public Health Nutr 40.
2018;21:3281–90. [32] Heitmann BL, Lissner L. Dietary underreporting by obese individuals
Is it spe-
[12] te Velde SJ, van Nassau F, Uijtdewilligen L, van Stralen MM, Cardon G, De cific or non-specific? BMJ 1995;311:986–9.
Craemer M, et al. Energy balance-related behaviours associated with over- [33] Te Morenga L, Mallard S, Mann J. Dietary sugars and body weight: systematic
weight and obesity in preschool children: a systematic review of prospective review and meta-analyses of randomised controlled trials and cohort studies.
studies. Obes Rev 2012;13:56–74. BMJ 2012;346:e7492.
[13] Faul F, Erdfelder E, Lang AG, Buchner AG. *Power 3: a flexible statistical power [34] Pfeiffer AFH, Keyhani-Nejad F. High glycemic index metabolic damage
A piv-
analysis program for the social, behavioral, and biomedical sciences. Behav otal role of GIP and GLP-1. Trends Endocrinol Metab 2018;29:289–99.
Res Methods 2007;39:175–91. [35] Schwingshackl L, Hoffmann G, Kalle-Uhlmann T, Arregui M, Buijsse B, Boeing
[14] Galobardes B, Shaw M, Lawlor DA, Lynch JW, Davey Smith G. Indicators of socio- H. Fruit and vegetable consumption and changes in anthropometric variables
economic position (part 1). J Epidemiol Community Health 2006;60:7–12. in adult populations: a systematic review and meta-analysis of prospective
[15] World Health Organization. Obesity. Available at: https://www.who.int/news- cohort studies. PloS One 2015;10:e0140846.
room/facts-in-pictures/detail/6-facts-on-obesity. Accessed December 1, 2021. [36] Howarth N, Saltzman E, Roberts S. Dietary fiber and weight regulation. Nutr
[16] World Health Organization. Waist circumference and waist
hip ratio: report Revs 2001;59:129–39.
of a WHO expert consultation. Geneva, Switzerland: World Health Organiza- [37] de Oliveira MC, Sichieri R. Venturim Mozzer R. A low-energy-dense diet adding
tion; 2008. p. 1–47. fruit reduces weight and energy intake in women. Appetite 2008;51:291–5.
[17] Anastasiou CA, Fappa E, Zachari K, Mavrogianni C, Van Stappen V, Kivela € J, et al. [38] Rodriguez-Casado A. The health potential of fruits and vegetables phytochemi-
Development and reliability of questionnaires for the assessment of diet and cals: notable Examples. Crit Rev Food Sci Nutr 2016;56:1097–107.
physical activity behaviors in a multi-country sample in Europe the Feel4Dia- [39] Davis HC. Can the gastrointestinal microbiota be modulated by dietary fibre to
betes Study. BMC Endocr Disord 2020;20:135. treat obesity? Ir J Med Sci 2018;187:393–402.
[18] Hemio € K, Po
€ lo
€nen A, Ahonen K, Kosola M, Viitasalo K, Lindstro € m J. A simple [40] Ruxton CHS, Myers M. Fruit juices: are they helpful or harmful? An evidence
tool for diet evaluation in primary health care: validation of a 16-item food review. Nutrients 2021;13:1815.
intake questionnaire. Int J Environ Res Public Health 2014;11:2683–97. [41] Charlton K, Kowal P, Soriano MM, Williams S, Banks E, Vo K, et al. Fruit and
[19] World Health Organization. Diet, nutrition and the prevention of chronic dis- vegetable intake and body mass index in a large sample of middle-aged Aus-
eases. Geneva, Switzerland: World Health Organization; 2003. tralian men and women. Nutrients 2014;6:2305–19.
[20] Bernstein AM, de Koning L, Flint AJ, Rexrode KM, Willett WC. Soda consumption [42] Eurostat. Daily consumption of fruit and vegetables by sex, age and body mass
and the risk of stroke in men and women. Am J Clin Nutr 2012;95:1190–9. index. Available at: https://ec.europa.eu/eurostat/databrowser/view/HLTH_E-
[21] Tabachnick BG, Fidell LS. Using multivariate statistics. 5th ed. Upper Saddle HIS_FV3M__custom_1775130/default/table?lang=en. Accessed December 1, 2021.
River, NJ: Pearson Allyn & Bacon; 2007. [43] World Health Organization. Increasing fruit and vegetable consumption to
[22] Thurstone LL. Multiple factor analysis: a development and expansion of vec- reduce the risk of noncommunicable diseases. Available at: http://www.who.
tors of the mind. Chicago, Illinois: University of Chicago; 1947. int/elena/titles/fruit_vegetables_ncds/en/. Accessed December 1, 2021.</bib>
[23] Kline P. An easy guide to factor analysis. London, UK: Routledge; 2002. T ag e dP [44] Eurostat. Overweight and obesity
BMI statistics. Available at: https://ec.
[24] Bland JM, Altman DG. Cronbach's alpha. BMJ 1997;314:572. europa.eu/eurostat/statistics-explained/index.php?title=Overweight_and_o-
[25] Safaei M, Sundararajan EA, Driss M, Boulila W, Shapi'i A. A systematic literature besity_-_BMI_statistics. Accessed December 1, 2021.
review on obesity: understanding the causes & consequences of obesity and [45] Darmon N, Drewnowski A. Contribution of food prices and diet cost to socio-
reviewing various machine learning approaches used to predict obesity. Com- economic disparities in diet quality and health: a systematic review and analy-
put Biol Med 2021;136:104754. sis. Nutr Revs 2015;73:643–60.
TagedP[26] Eurostat. Over half of adults in the EU are overweight. Available at: https://ec. [46] Sørensen K, Pelikan JM, Ro € thlin F, Ganahl K, Slonska Z, Doyle G, et al. Health lit-
europa.eu/eurostat/web/products-eurostat-news/-/ddn-20210721
2. eracy in Europe: comparative results of the European health literacy survey
Accessed December 1, 2021. (HLS-EU). Eur J Public Health 2015;25:1053–8.
[27] Qin P, Li Q, Zhao Y, Chen Q, Sun X, Liu Y, et al. Sugar and artificially sweetened [47] Yumuk V, Tsigos C, Fried M, Schindler K, Busetto L, Micic D, et al. European
beverages and risk of obesity, type 2 diabetes mellitus, hypertension, and all- guidelines for obesity management in adults. Obes Facts 2015;8:402–24.
cause mortality: a dose-response meta-analysis of prospective cohort studies. [48] Garber CE. The health benefits of exercise in overweight and obese patients.
Eur J Epidemiol 2020;35:655–71. Curr Sports Med Rep 2019;18:287–91.
[28] Chazelas E, Debras C, Srour B, Fezeu LK, Julia C, Hercberg S, et al. Sugary drinks, [49] Schlesinger S, Neuenschwander M, Schwedhelm C, Hoffmann G, Bechthold A,
artificially-sweetened beverages, and cardiovascular disease in the NutriNet- Boeing H, et al. Food groups and risk of overweight, obesity, and weight gain:
Sante
Cohort. J Am Coll Cardiol 2020;76:2175–7. a systematic review and dose
response meta-analysis of prospective studies.
[29] Rother KI, Conway EM, Sylvetsky AC. How non-nutritive sweeteners influence Adv Nutr 2019;10:205–18.
hormones and health. Trends Endocrinol Metab 2018;29:455–67. [50] Butler L, Popkin BM, Poti JM. Associations of alcoholic beverage consumption
[30] Houchins JA, Burgess JR, Campbell WW, Daniel JR, Ferruzzi MG, McCabe GP, with dietary intake, waist circumference, and body mass index in U.S. adults:
et al. Beverage versus solid fruits and vegetables: effects on energy intake and National Health and Nutrition Examination Survey 2003
2012. J Acad Nutr
body weight. Obesity 2012;20:1844–50. Diet 2018;118:409–20. e3.